7.6 Pervasive problem: Example post-treatment bias

  • All empirical papers, top 3 political science journals, 2010-2015
    • “40% explicitly conditioned on a post treatment variable […] 27% conditioned on a variable that could plausibly be posttreatment […] 33% […] no post-treatment variables included in their analyses […] two-thirds […] that make causal claims condition on post treatment variables.” (Acharya, Blackwell, and Sen 2015, 1)
  • Q: How about sociology?

References

Acharya, Avidit, Matthew Blackwell, and Maya Sen. 2015. “Explaining Causal Findings Without Bias: Detecting and Assessing Direct Effects.” The American Political Science Review.